The Future of Computing
Articles,  Blog

The Future of Computing

My name is Christof Teuscher. I’m a professor in the
department of electrical and computer engineering at
Portland State University. So, machine learning is, if
you go back 50 years or so, the challenge was to
build machines that can do things that humans can’t do
or that humans can do well like playing chess or something like that. These are solved problems,
but in those days machine learning is more
about pattern recognition and doing things in a
smart and adaptive way where human intelligence basically fails or basically our brain is too small to solve these problems, right? Maybe a good example is
airline reservations systems where you can’t do it as a
human. It’s just to complicated. There’s too many variables. There’s too many things
that you have to consider. Our machines are doing that
job really, really well but we’re looking at tasks
where humans used to be good like face recognition, things like that, and machines weren’t and that is changing. I think now we have machines that are more reliable
at recognizing faces and can do things that we
as humans used to be good but now machines are taking over, so it’s a whole new, exciting field. We’re looking at the, what
we call, the computing stack that goes, basically from
the hardware of the devices that do the computation
up to the application as sort of as an all-encompassing problem. You can’t really solve issues
as one of these levels only in these days. It has to really address all the different abstraction levels and the different components
that you’re going to work with so it’s very inter-disciplinary
for that specific reason and involves computer science and it involves electrical engineering, but it also involves biology if you’re changing this up straight and are computing with biomolecules instead of transistors, for example. So, we’re very much interested in using novel types of devices, building computing
archetectures with those and then also finding new
ways of programming them and very often in these
days we use machine-learning as a way to program these
systems and make them do things that we cannot do otherwise,
so learning adaptation, these are very interesting properties that we’re interested in. That goes in the direction,
the generally hot topic in these days is machine learning and AI and, you know, machines
doing things that we, as humans are not
actually so good at doing. But we also have more exotic things like a biomolecular computation project where the basic building
blocks are not transistors or memristers, it’s biomolicules,
DNA strands that combine, recombine and so, computing
with those sort of things is a whole other level of
challenges and it has applications for biomedical and dynamic fields. Smart, adaptive devices that
you could potentially swallow, wear on your skin, under your skin that monitor your blood
sugar, whatever it is so there’s lots of
really cool applications in the biomedical field. We’re going to need machines that adapt, that learn, that adapt to
your lifestyle, for example. That’s another research
axis as an example. Moving into no-man’s land,
high-risk, high-payoff. That’s, I would say is the
summary of the research that we’re doing. If things work, great,
but there’s high-risk and then you have the
avenues that we’re exploring may not work, but that’s part of research. It’s never straight line and we really have to
explore that no-man’s land to find that well, what is
a potential avenue forward? My name is Christof Teuscher. Our research purpose is to build smarter and better computers to
make this a better world.

Leave a Reply

Your email address will not be published. Required fields are marked *